Signatures of adaptive evolution in the genetic sequences of human pathogenic RNA viruses

Abstract

Rapid evolution of human pathogenic RNA viruses can undermine our efforts to control infection and transmission. For instance, seasonal influenza undergoes adaptive evolution directed by selection to evade antibodies, resulting in continual antigenic changes and ne- cessitating nearly annual vaccine updates to match the circulating viruses. Additionally, influenza’s propensity to undergo adaptive evolution presents another challenge for vaccine production as it adapts to the eggs the vaccine strain is grown in, often altering antigenic- ity and impacting vaccine effectiveness. Adaptive evolution leaves certain marks on the genome, which can be identified and interpreted through phylogenetic and sequence-based analyses. In this dissertation, I employ a variety of computational techniques to find and interpret these marks in influenza H3N2 and coronaviruses. First, I systematically identify H3N2 mutations that adapt the virus to replication in eggs, show that epistatic interactions between these mutations constrain the adaptive evolution of H3N2, and describe the po- tential antigenic impact of these egg-adapted mutations. While influenza’s adaptive (and particularly antigenic) evolution is widely-appreciated, it is not as well understood which other RNA viruses undergo similar evolution. Thus, I next utilize several complementary methods to show evidence of recurrent adaptive evolution in seasonal coronaviruses that is localized to the viral gene targeted by human antibodies. Finally, I use novel methods to comprehensively scan the genome of SARS-CoV-2 for evidence of adaptive evolution, identify specific adaptive mutations, and show temporal structure to the evolution of SARS-CoV-2 during the first year and a half of the pandemic. Together, the work in this dissertation demonstrates how genetic sequence data can be used to understand the adaptive evolution of human pathogenic RNA viruses, which informs how these viruses can be most effectively controlled.